Every race read before post — our confidence tier, the coverage pool with a one-line note on each horse, and the overlays. Graded on The Record once they run.
LOCK = single-friendly · CHAOS = wide-open, spread the pool. 🏆 is our top single. The % is the morning-line implied win chance (the market's number, normalized) — the pool is sized to cover ~70/80/85% of it.
PX (members, experimental) scores each runner's last race on today's surface and trip type: how close it was at the 2nd call, its pace figure against today's field, ground gained to the wire (allowing for where it sat), and how fast that race closed. Q5–Q1 are fifths set by a fixed cut from the backtest (so a card can hold more or fewer than one in five); Q5 is the best. A — means no comparable last race (debut, first route, first turf) — unscored, not a negative, and the PX filters dim those runners too. In testing, Q5 won about 1.3× as often as its morning line implied and Q1 longshots about half as often — but the final odds already price it, so it is a morning-line edge for building pools, not a reason to take a short price.
MP (members, experimental) is our model's win probability, shown under the morning line's Win%: a frozen model that starts from the M/L and adjusts it with the form cycle (recent pace figures, days off, class move, last-race odds) and raw handicapping inputs. It is better calibrated than the morning line in testing, but the final odds already price it, and it orders the top three contenders worse than the M/L, so it never re-ranks our picks. Green or red means it differs from the M/L by 2+ points. 4th marks the best MP outside the M/L top three: in backtests, using it as the 4th horse in a leg covered the winner slightly more often. Graded live on the Record page.
The % beside each jockey and trainer is their win rate in this race's class type (e.g. Maiden, Stakes) — hover it for the sample size.
What each flag is telling you at a glance — the colored tag is the read. Full detail & the math →